Except websites can tell what base OS you run using browser fingerprinting. It os impossible to lie aboit your OS because of the differences in platforms.
That isnt a great defense against malware “imho”. Security through assuming the threat actor is lazy is just not security. It doesnt take like any effort on their part to just use some off-the-shelf OS fingerprinting code. It isnt worth it either because it contributes to your overall fingerprint, since normal RFP users have a standardized useragent for Windows and Linux separately.
It really isnt any defense. All a website can do is initiate a download, websites are sandboxed by default. You still have to run the executable, which doesnt really apply to Linux because the file will have no executable permission.
Security is layers, i utilize apparmor and firejail personally. And in fact 90% of widespread malware specifically relies on lazy people. Often targets default passwords etc
If you are relying on community sandboxing profiles and not making your own, i can understand why Firejail is interesting as a choice because of its large community.
If you are making your own, consider checking out Bubblewrap (available on most Linux systems), Bubblejail), Crablock, and Sydbox, which all use unprivileged sandboxes.
You can lie, but that doesnt mean that a website cant still tell your base OS if they use JS platform fingerprinting. Arkenfox, the base config which Librewolf is based off of says the exact same thing. Go to CreepJS and see it get your platform regardless.
Secondly, do you block all JS? NoScript is not a silver bullet and doesnt stop fingerprinting, it is itself identified by the CreepJS test site. It may in this case reduce the chance of OS fingerprinting, but pure CSS methods exist as well.
Additionally, NoScript is laregly redundant with uBlock Origin since you can do everything that it offers, such as blocking 3rd party scripts/iframes/all, block fonts, block JS, and it is very granular.
Generally browser fingerprinting is used to identify individual browser sessions across IP addresses. This mostly takes into account reported features and capabilities of the browser and OS to the website. Fingerprinting isn’t looking for specific info your browser reports, it’s taking it all and hashing it to get a unique id specific to the browser. Because it’s hashed, it can’t be reversed to identify the OS from the hash.
Sure a malicious website could Ignore the user agent and probe for some hardware capabilities that are specific to Linux, but that would be a lot of effort to probe various things which are set differently across all different browsers. I can’t speak for bad actors, but I wouldn’t spend the effort to check if the user agent is spoofed, if 95% of the time it’s accurate to get the OS type.
It is trivial to identify OS platform because browser work differently on each platform. Wjat Librewolf does with useragent on Linux actually is makes users stand out more because it isn’t what privacy.resistFingerprinting (RFP) reports on normally.
Hackers (like the comment scenario i was responding to) are substantially more likely to employ platform fingerprint than trust a fale useragent. And loads general websites employ fingerprinting, meaning deviation from default RFP behaviour makes you stand out (more than you already do by using RFP since it is a small pool already).
Agreed, I’m not saying it’s impossible to detect the OS, but it’s even more trivial for an adversary to regex the User Agent and serve the malware for that OS. The average user doesn’t even know what a User Agent is, and that’s who the drive by malware websites are counting on to infect because they’re easy targets.
Just like a real fingerprint, that will only identify the fingerprint to a person, not tell you that the fingerprint is from someone who is European. Fingerprints are used to track you across different websites, and build a profile of you for advertising.
My logic was that it is much more likely that someone will spoof there useragent already if they are on Linux. If threat actor is targeting not just Windows but also Linux, they probably would understand the very real likelyhood of platform spoofing.
Twitch.TV will tell you that you need to use a supported browser if you connect with Linux in your user agent no matter what browser you use. Changing Linux to Windows in your agent with no other changes resolves this issue.
Sure, you can lie to get around restrictions that only care about useragent. But it is still clear as day what platform you are using, even against basic threat actors. It is not “anti-malware” and it reduces the pool of users you blend in with. Firefox’s privacy.resistFingerprinting preference uses the anti-fingerprinting strategy of “normalize metrics to create a crowd” (inherited from Tor browser), but this is on a per-OS basis because it is impossible to effectively lie about your platform (ei. a pool of Windows RFP users, MacOS, and BSD, etc).
Except websites can tell what base OS you run using browser fingerprinting. It os impossible to lie aboit your OS because of the differences in platforms.
Sure, if they put in the effort. Most don’t even when Distributing malware
That isnt a great defense against malware “imho”. Security through assuming the threat actor is lazy is just not security. It doesnt take like any effort on their part to just use some off-the-shelf OS fingerprinting code. It isnt worth it either because it contributes to your overall fingerprint, since normal RFP users have a standardized useragent for Windows and Linux separately.
Nobody said it was a great defense.
It really isnt any defense. All a website can do is initiate a download, websites are sandboxed by default. You still have to run the executable, which doesnt really apply to Linux because the file will have no executable permission.
Buddy, it’s not that deep.
Security is layers, i utilize apparmor and firejail personally. And in fact 90% of widespread malware specifically relies on lazy people. Often targets default passwords etc
Firejail is a large SETUID binary which can (and has) aid in privilege escalation. It is recommended to avoid it for this reason.
See: https://madaidans-insecurities.github.io/linux.html#firejail
If you are relying on community sandboxing profiles and not making your own, i can understand why Firejail is interesting as a choice because of its large community.
If you are making your own, consider checking out Bubblewrap (available on most Linux systems), Bubblejail), Crablock, and Sydbox, which all use unprivileged sandboxes.
You can lie about your fingerprint very much in fact it is the default on librewolf
You can lie, but that doesnt mean that a website cant still tell your base OS if they use JS platform fingerprinting. Arkenfox, the base config which Librewolf is based off of says the exact same thing. Go to CreepJS and see it get your platform regardless.
You sweet summer child I use noscript
Firstly there is no need to be condescending.
Secondly, do you block all JS? NoScript is not a silver bullet and doesnt stop fingerprinting, it is itself identified by the CreepJS test site. It may in this case reduce the chance of OS fingerprinting, but pure CSS methods exist as well.
Additionally, NoScript is laregly redundant with uBlock Origin since you can do everything that it offers, such as blocking 3rd party scripts/iframes/all, block fonts, block JS, and it is very granular.
Bottom line, you are fingerpintable.
Generally browser fingerprinting is used to identify individual browser sessions across IP addresses. This mostly takes into account reported features and capabilities of the browser and OS to the website. Fingerprinting isn’t looking for specific info your browser reports, it’s taking it all and hashing it to get a unique id specific to the browser. Because it’s hashed, it can’t be reversed to identify the OS from the hash.
Sure a malicious website could Ignore the user agent and probe for some hardware capabilities that are specific to Linux, but that would be a lot of effort to probe various things which are set differently across all different browsers. I can’t speak for bad actors, but I wouldn’t spend the effort to check if the user agent is spoofed, if 95% of the time it’s accurate to get the OS type.
It is trivial to identify OS platform because browser work differently on each platform. Wjat Librewolf does with useragent on Linux actually is makes users stand out more because it isn’t what privacy.resistFingerprinting (RFP) reports on normally.
Hackers (like the comment scenario i was responding to) are substantially more likely to employ platform fingerprint than trust a fale useragent. And loads general websites employ fingerprinting, meaning deviation from default RFP behaviour makes you stand out (more than you already do by using RFP since it is a small pool already).
Agreed, I’m not saying it’s impossible to detect the OS, but it’s even more trivial for an adversary to regex the User Agent and serve the malware for that OS. The average user doesn’t even know what a User Agent is, and that’s who the drive by malware websites are counting on to infect because they’re easy targets.
Just like a real fingerprint, that will only identify the fingerprint to a person, not tell you that the fingerprint is from someone who is European. Fingerprints are used to track you across different websites, and build a profile of you for advertising.
Yeah okay.
My logic was that it is much more likely that someone will spoof there useragent already if they are on Linux. If threat actor is targeting not just Windows but also Linux, they probably would understand the very real likelyhood of platform spoofing.
Twitch.TV will tell you that you need to use a supported browser if you connect with Linux in your user agent no matter what browser you use. Changing Linux to Windows in your agent with no other changes resolves this issue.
Worked on my machine last I checked, and I don’t have a useragent spoofer
Sure, you can lie to get around restrictions that only care about useragent. But it is still clear as day what platform you are using, even against basic threat actors. It is not “anti-malware” and it reduces the pool of users you blend in with. Firefox’s privacy.resistFingerprinting preference uses the anti-fingerprinting strategy of “normalize metrics to create a crowd” (inherited from Tor browser), but this is on a per-OS basis because it is impossible to effectively lie about your platform (ei. a pool of Windows RFP users, MacOS, and BSD, etc).